Towards Spatio-Temporal Extrapolation of Phase-Field Simulations with Convolution-Only Neural Networks

Christophe Bonneville, Nathan Bieberdorf, Pieterjan Robbe, Mark Asta, Habib Najm, Laurent Capolungo, Cosmin Safta

arXivarxiv/2601.04510Jan 8, 2026cs.CEscore 0

Phase-field simulations of liquid metal dealloying (LMD) can capture complex microstructural evolutions but can be prohibitively expensive for large domains and long time horizons. In this paper, we introduce a fully convolutional, conditionally parameterized U-Net surrogate designed to extrapolate far beyond its training data in both space and time. The architecture integrates convolutional self-attention, physically informed padding, and a flood-fill corrector method to maintain accuracy under extreme extrapolation, while conditioning on simulation parameters allows for flexible time-step…

SourceJSON

Comments and reviews

No comments yet.

API

curl /api/v1/records/arxiv/2601.04510
curl -X POST /api/v1/records/arxiv/2601.04510/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2601.04510/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"body": "…", "review": true}'